Clinical dialogue transcription error correction using Seq2Seq models.

Nanayakkara, Gayani; Wiratunga, Nirmalie; Corsar, David; Martin, Kyle; Wijekoon, Anjana

Abstract

Good communication is critical to good healthcare. Clinical dialogue is a conversation between health practitioners and their patients, with the explicit goal of obtaining and sharing medical information. This information contributes to medical decision-making regarding the patient and plays a crucial role in their healthcare journey. The reliance on note taking and manual scribing processes are extremely inefficient and leads to manual transcription errors when digitizing notes. Automatic Speech Recognition (ASR) plays a significant role in speech-to-text applications, and can be directly used as a text generator in conversational applications. However, recording clinical dialogue presents a number of general and domain-specific challenges. In this paper, we present a seq2seq learning approach for ASR transcription error correction of clinical dialogues. We introduce a new Gastrointestinal Clinical Dialogue (GCD) Dataset which was gathered by healthcare professionals from a NHS Inflammatory Bowel Disease clinic and use this in a comparative study with four commercial ASR systems. Using self-supervision strategies, we fine-tune a seq2seq model on a mask-filling task using a domain-specific PubMed dataset which we have shared publicly for future research. The BART model fine-tuned for mask-filling was able to correct transcription errors and achieve lower word error rates for three out of four commercial ASR outputs.

Citation

NANAYAKKARA, G., WIRATURNGA, N., CORSAR, D., MARTIN, K. and WIJEKOON, A. 2022. Clinical dialogue transcription error correction using Seq2Seq models. In Shaban-Nejad, A., Michalowski, M. and Bianco, S. (eds.) Multimodal AI in healthcare: a paradigm shift in health intelligence; selected papers from the 6th International workshop on health intelligence (W3PHIAI-22), co-located with the 34th AAAI (Association for the Advancement of Artificial Intelligence) Innovative applications of artificial intelligence (IAAI-22), 28 February - 1 March 2022, [virtual event]. Studies in computational intelligence, 1060. Cham: Springer [online], pages 41-57. Available from: https://doi.org/10.1007/978-3-031-14771-5_4

• Presentation Conference Type: Conference Paper (published)
• Conference Name: 6th International workshop on health intelligence (W3PHIAI-22), co-located with the AAAI (Association for the Advancement of Artificial Intelligence) 34th Innovative applications of artificial intelligence (IAAI-22)
• Start Date: Feb 28, 2022
• End Date: Mar 1, 2022
• Acceptance Date: Dec 3, 2021
• Online Publication Date: Nov 29, 2022
• Publication Date: Dec 31, 2022
• Deposit Date: Oct 25, 2022
• Publicly Available Date: Nov 30, 2023
• Publisher: Springer
• Peer Reviewed: Peer Reviewed
• Pages: 41-57
• Series Title: Studies in computational intelligence (SCI)
• Series Number: 1060
• Series ISSN: 1860-949X; 1860-9503
• Book Title: Multimodal AI in healthcare: a paradigm shift in health intelligence.
• ISBN: 9783031147708
• DOI: https://doi.org/10.1007/978-3-031-14771-5_4
• Keywords: Clinical dialogue transcription; Automatic speech recognition; Error correction
• Public URL: https://rgu-repository.worktribe.com/output/1686809
• Related Public URLs: https://rgu-repository.worktribe.com/output/1686647
• Additional Information: A pre-print version of this article was first available as: NANAYAKKARA, G., WIRATUNGA, N., CORSAR, D., MARTIN, K. and WIJEKOON, A. 2022. Clinical dialogue transcription error correction using Seq2Seq models. Hosted on arXiv [online]. Available from: ; https://doi.org/10.48550/arXiv.2205.13572